POLICY TRADE-OFFS ELECTRICITY SYSTEM

Transcription

1 Prof. Dr. Johan Albrecht Ruben Laleman September, 2014 POLICY TRADE-OFFS for the BELGIAN ELECTRICITY SYSTEM

2 II

3 ABSTRACT The outlook of the Belgian electricity system is increasingly unpredictable and challenging. Belgium is confronted with a nuclear phase out in a liberalized European electricity market which is strongly impacted by climate and renewable energy policies. In the market context of today, incentive schemes focus on renewable energy sources which are sheltered from market dynamics. Load factors of conventional power plants have dropped markedly due to a stagnating demand and an increasing share of intermittent renewable electricity generation with low variable costs. As a consequence, the investment climate for controllable, non-intermittent assets is very problematic. In this report we evaluate the expected changes in the Belgian electricity supply from 2014 until We also estimate cost implications of the nuclear phase out, combined with a decrease in old fossil capacity and an increase in renewable electricity generation. In a baseline scenario we find that, in the near future, Belgium strongly will have to rely on electricity imports to meet peak demand. This import dependency can eventually increase black out risks. We assume that this risk is unacceptable for policymakers and therefore assess several secure supply scenarios for the future of the Belgian electricity system. We assume that a reserve margin of at least 5% always needs to be maintained, and evaluate the implications of meeting this benchmark. Incentives for capacity growth, system flexibility and system reliability are compared. We consider investments in new assets, the prolonged use of old thermal assets and demand-side measures. For various scenarios we estimate subsidy costs, overall system costs and related outcomes such as surplus problems. In the sets of scenarios, we compare two options to deal with the variability of renewable generation technologies. The first option is similar to the current situation and assumes that renewables have grid priority as well as priority in the merit order due to their output related support (production subsidies). The second option assumes that renewables are obliged to participate in the market to a certain extent. Finally, we combine all scenarios to estimate the costs of a balanced approach. Contact Information: Prof. Dr. Johan Albrecht, Universiteit Gent - Faculteit Economie & Bedrijfskunde Tweekerkenstraat 2, B-9000 Gent Johan.Albrecht@ugent.be / Tel: ++ (0) / Fax: ++ (0) III

4 EXECUTIVE SUMMARY The Belgian electricity mix based on nuclear and fossil fuels will evolve into a system with an increasing share of renewable energy technologies. The Belgian situation is very special because a large share of the existing capacity is planned to be phased out or closed in the next 15 years. This decline in old nuclear and fossil capacity can already in the very near future ( ) results in shortages, especially at very cold winter evenings. Current policies trigger mainly investments in (intermittent) renewable capacity. Incentives for investments in system reliability or system flexibility are currently lacking. Hence we face the deployment of solar and wind technologies without new investments in controllable assets or firm capacity. This transition is challenging from a system perspective. In this report various scenarios are put forward to provide incentives for both renewable and controllable assets. More specifically, we foresee a support system for wind, PV and biomass to promote renewable energy technologies (and meet European targets). In addition, remunerations for the availability of capacity are introduced to support gas fired (OCGT, CCGT) and biomass combustion technologies. In a country like Belgium, biomass power plants provide the only large scale firm and renewable capacity. In theory, hydro power could also serve this purpose but Belgium has very limited potential to increase its hydro capacity. Other technologies (CCS, wave and tidal energy...) which are not yet commercially viable are not considered in this study. In total, 16 scenarios have been evaluated in this report. Half of these assume moderate growth of installed renewable capacities (business as usual or BAU RES scenarios); the other 8 assume strong renewables policies ( High RES scenarios ). The scenarios vary in two other dimensions as well. One dimension is the degree of market participation that is expected from the renewable technologies. The other dimension is the support system to facilitate investments in firm capacity and to meet peak demand. For all scenarios a minimal reserve margin of 5% was set as a policy goal, in order to have a sufficiently secure electricity system. For all of these 16 scenarios we have calculated the cumulative additional subsidy costs - additional to the subsidies already in place today - and total annual as well as cumulative system costs for period Also, an assessment of problematic volumes of electricity oversupply was added. When comparing all the scenarios evaluated in this study, we found that the least cost option to meet the 5% reserve margin requirement is a scenario with a moderate growth of total renewable capacity combined with a modest degree of market participation by the renewable assets. This type of market participation mainly consists of flexible use of biomass plants and an option to curtail PV and wind at times of low demand combined with favourable weather conditions. The lowest cost scenario also assumes an increased use of demand response measures ( Demand Side Management ) to postpone or replace the need for firm capacity (OCGT, CCGT or eventually biomass). An additional way to dampen costs is to keep old assets on line as back-up for cold winter peaks. IV

5 The estimated total cumulative (undiscounted) subsidy cost for this least cost generation scenario is 21 Billion for the period The latter subsidy cost only contains financial support for generation assets. In contrast, the total system cost also contains the costs of other generation assets (used without new subsidies) and the costs of transmission and distribution. With the lowest cost scenario, the cumulative system costs in the period amount to about 160 Billion. Total annual system costs will increase from 6 Billion (assuming the use of amortised nuclear assets) in 2014 to roughly 10 Billion in With this scenario, the share of renewable electricity in total domestic supply in 2030 will be about 27%. This is likely to be sufficient to meet European RES-targets for 2030 (still under debate). The energy transition up to 2030 will not be a cheap transformation. Total cumulative subsidy costs ( ) for the generation scenarios range from 21 Billion to 41 Billion and cumulative system cost are in the range of 160 to 180 Billion. As a consequence, appropriate policy choices to minimize the cumulative cost of energy security can be 20 Billion less expensive between 2014 and 2030 than the most expensive policy options. Total annual system costs in 2030 range from 9.9 Billion (low share of renewables, flexible electricity system) to 11.6 Billion (high share of renewables, inflexible system). The most expensive scenario consists of a system with a high share of intermittent renewables, inflexible biomass plants, a lot of (new and efficient) gas-fired power plants with very low load factors and a limited contribution of demand response measures. The share of renewable generation will vary between 27% and 57% by Incentives for renewable generation will need to change from a current production based perspective to a more system-wide perspective with system flexibility and reliability needs taken into account. A new incentive scheme for renewables is essential because production subsidies hamper the further expansion of new renewable capacity, due to risks for oversupply, grid instability as well as high system-based costs. V

6 GLOSSARY General RES Renewable Energy Sources ETS Emissions Trading System CCGT Combined Cycle Gas Turbine OCGT Open Cycle Gas Turbine CHP Combined Heat and Power PV Photovoltaic (Panels) - Solar Panels LF Load Factor DSM Demand Side Management CWE Central West European (electricity trading region) O&M Operation and Maintenance GHG Greenhouse Gasses (CO2, CH4, ) RM Reserve Margin LCOE Levelized Cost of Electricity CF Capacity Factor Scenarios BAU Business as Usual (BAU RES scenario) CFD Contract For Differences scenario (p 19) CFD-MP CFD-Market Participation scenario (shedding of RES, p 20) OT Old Thermal (incentives for old fossil capacity, p 21) New Incentives for NEW fossil capacity (p 21) DSM DSM- Scenario: strong focus on DSM deployment (p 22) MP-IR Market Participation - Intermittent RES (p 34) Companies, public and private organizations ENTSO-E European Network of Transmission System Operators - Electricity CREG Commission for the Regulation of Electricity and Gas markets (Belgium) DENA Deutsche Energie-Agentur GmbH (German Energy Agency) DG-Energy Directorate-General for Energy (EU commission) FOD-Economy Federal Government Agency - Economy (Belgium) NEA Nuclear Energy Agency IEA International Energy Agency OECD Organization for Economic Co-operation and Development UBS UBS- AG Swiss Bank US-PJM United States Regional Transmission Organization (Pennsylvania, New Jersey, ) EREC European Renewable Energy Council EPIA European Photovoltaic Industry Association EWEA European Wind Energy Association Units MW Mega Watt (10 8 W) TWh Tera Watt hours (10 12 Wh) GWh Giga Watt hours (10 9 Wh) VI

7 VII

8 Table of Contents ABSTRACT... III EXECUTIVE SUMMARY... IV GLOSSARY... VI 1 Introduction The Belgian Situation Current trends Future of electricity production - Plan Wathelet The Central West European electricity market Future electricity supply scenarios for Belgium Introduction Assumptions for all scenarios Evolution of peak demand Evolution of the carbon price Discount rate Cost calculations Technology assumptions Endogenous price model Distribution and transmission cost estimates Security of supply scenarios Business as usual (BAU) and high renewable scenario (High RES) Incentive scenarios to guarantee security of supply Introduction Renewable supply measures Incentives for firm capacity Results Electricity Supply in Annual and cumulative subsidy costs from 2014 to System costs from 2014 to Theoretical perspective VIII

9 4.3.2 Practical perspective Discussion Surplus Risk Method Results Surplus risk Surplus risk Surplus risk Surplus risk Discussion Alternative Scenario: Nuclear prolongation Conclusion Bibliography IX

10 1 Introduction More than ever, electricity markets in Europe are facing rapid changes due to a combination of many unprecedented challenges. First, there are the European 20/20/20 targets that aim to increase the share of renewable energy sources (RES) in total energy consumption (up to 20%), decrease greenhouse gas emissions (by 20%) and improve the overall energy efficiency of the European economy by the year Meanwhile, new targets for 2030 are on their way, but are still heavily debated. Also, a long and persistent economic crisis is increasing uncertainty and many member states are urged to reduce their deficits. On top of this, the recent accident in Fukushima has resulted in the rapid phase out of nuclear power plants in Germany. In Belgium and the U.K., old nuclear power plants are planned to retire in the next decades. In addition to these phase out plans, many other old assets in Europe will be closed down in the next decade. Cheap coal and the strong expansion of subsidized renewables with very low marginal costs are putting pressure on wholesale prices and pushing new and efficient gas plants (CCGT s) out of the market. Very low ETS prices are further complicating the situation for gas fired plants in Europe. New investments in non-intermittent, controllable assets are virtually non-existent, resulting in an increased fear of electricity shortages in some regions. More interconnection can help solve some issues, but local opposition against large infrastructure projects and complex licensing procedures complicate rapid development. In this complex world governments (local, federal and European) are urged to come up with measures to provide a more attractive and secure investment climate for energy (technology) companies. This report aims to pinpoint the problems occurring in the Belgian and the Central-West-European (CWE) electricity systems. We focus on the big impact of the Belgian nuclear phase out in combination with increasing (intermittent) renewable generation and the low profitability of traditional gas plants (CCGT s). Some future policy scenarios are presented and evaluated. We focus on Belgium, but keep the larger European landscape in mind. 1

11 Installed Capacity (MW) Johan Albrecht, Ruben Laleman 2 The Belgian Situation 2.1 Current trends The installed capacities of the various energy technologies in Belgium are presented in Figure 1 (Eurelectric, 2013). The share of renewables has increased markedly in the recent 5 years. However, nuclear capacity is still very dominant in the electricity production park with a share of about 30% in total capacity and about 40% of the controllable or firm capacity (total capacity without PV, hydro and wind). Despite the increase in total capacity, the sum of firm capacity has remained fairly constant at about MW. According to current government policies, the share of nuclear capacity will decrease drastically in the next 20 years. By from 2025 onwards, nuclear capacity (currently +/ MW) will be completely phased out. Meanwhile, several fossil fuelled plants will also be closed because of end of life or due to steep losses in profitability. The Belgian electricity landscape will therefore change dramatically. Figure 1: Share of Technologies in Belgian Electricity Capacity ( ) source: (Eurelectric, 2013) Hydro Wind PV Biomass + other RES Pumped hydro Fossil (other) Gas Coal Nuclear Firm Capacity Table 1 shows some recent load factors for the various technologies in the Belgian production park. A remarkable fact is that the load factors of the gas plants (in bold) decreased significantly since the economic crisis, while the load factor of the coals plants increased. This is caused by the combination of low CO 2 and low coal prices, increasing gas prices and a rather flat or even decreasing electricity demand. The drop in the nuclear load factors is due to the safety issues at the Doel 3 and Tihange 2 reactors, which were shut down in Their start-up took place in June We can therefore assume that the LF of nuclear capacity at the end of 2013 did return to normal values. Due to the large variation in load factors, the technologies share in electricity production is very different from the share in total capacity (Figure 2 vs. Figure 1). Renewable energy technologies may have a relatively high share in total capacity, their share in electricity generation is still quite modest. Nevertheless, the share of renewable electricity in domestic production is rising steadily - mainly because of PV growth - to attain 14% in

12 Production (TWh) Share Res (%) Johan Albrecht, Ruben Laleman Table 1: Load factor of various technologies ( ) (Eurelectric, 2013) Load Factor (%) Nuclear 88% 88% 74% Fossil (other) 14% 14% 11% Coal 74% 70% 91% Gas 58% 41% 34% Biomass + other RES 76% 58% 62% Hydro 29% 19% 37% Wind 16% 25% 22% PV 7% 10% 7% Pumped Hydro 12% 11% 11% Average 60% 52% 43% Figure 2: Share in Domestic Production ( ), excluding import and export, source: (Eurelectric, 2013) % 14% 12% 10% 8% 6% 4% 2% 0% Pumped Hydro PV Wind Hydro BM + others Gas Coal Fossil (other) Nuclear Share Res Figure 2 shows that domestic gross electricity production (TWh) has decreased steadily in recent years. This is largely due to a strong increase of imported electricity. It is therefore interesting to take a look at total gross and net consumption as well, to see how much electricity is imported and how much is actually consumed in Belgium. In order to obtain the annual net consumption of electricity we need to take import 1 and losses into account. We obtain gross production of electricity (red line in Figure 3) by adding the total amount of electricity produced domestically (Figure 2) and the balance of imported and exported electricity. By subtracting losses and internal electricity consumption (electricity used by the power plant) from gross electricity production (+/- import/export) we obtain net electricity consumption (red line in Figure 3). Figure 3 shows that since the financial and economic crisis - which started in the average annual consumption of electricity in Belgium has dropped by about 7% (compared to ). The drop in 1 Electricity is imported and exported on a daily basis, depending on electricity prices. In recent years we have become net-importers. Over the whole year, more is imported then exported. 3

13 Peak Load (MW) Electricity Consumption/Production (GWh) Johan Albrecht, Ruben Laleman total demand does not necessarily result in a higher security of supply. It is the (instantaneous) peak demand that determines whether or not a country faces a high or low risk for a black out. Figure 3: Annual gross electricity production (+ net import) and consumption (source: (IEA, 2013)) Gross Production + Net Import Net Consumption Data for overall peak demand are hard to find because of auto-production. Peak load, on the other hand, is measurable and precise figures are available. However, it is important to stress that peak demand (including auto-production) is higher than peak load (= electricity taken from the grid) so we are underestimating peak demand when analyzing peak load. Nevertheless, we observe that peak load did not decrease as dramatically as the total annual demand. Peak load reached a maximum of about MW in 2007, to drop to around MW in 2012 (Figure 4). Whether or not this evolution resulted in a lower black out risk depends on the evolution of available firm capacity. It is very unclear whether this decrease in peak load is a structural trend or whether the decline is due to the economic crisis. Another explanation relates to the widening of the gap between peak load and peak demand. Overall, there is no certainty that the peak load (or peak demand) will remain low in the coming years. Figure 4: Peak Load (LOAD; source: (Elia, 2013) - remark: auto-production not included) Peak Load Lineair (Peak Load)

14 Available Nuclear Capacity (MW) Johan Albrecht, Ruben Laleman 2.2 Future of electricity production - Plan Wathelet In the summer of 2013 the Federal Minister for Energy, Environment, Mobility and State Reform, Mr. Wathelet, was able to get approval for his plan for the future of the Belgian electricity system. The Plan Wathelet foresees in the phase out of the two oldest nuclear reactors in 2015 (Doel 1 and Doel 2). A reactor in Tihange (Tihange 1) - that initially also was foreseen to be phased out in will be refurbished and its lifetime will be extended with another ten years 2. Thus, the short term decrease in nuclear capacity is not as radical as initially scheduled. However, the extension of Tihange 1 results in a stronger decrease in nuclear output in the period , when MW of nuclear capacity is scheduled to be phased out (Figure 5). Whether this will happen is at the moment unclear 3. Figure 5: Nuclear Phase out according to the "plan Wathelet" Prolonging the lifetime of Tihange 1 should result in fewer concerns about security of supply in Belgium on the short run. However, the nuclear phase out is not the only issue when it comes to security of supply. Many fossil fired plants are facing closure due to their age (old coal plants) or due to the fact that they are no longer profitable (as illustrated by the recent mothballing of recent gas plants). Also, increased capacities of wind and solar technologies come with a different kind of security of supply problem. As their variable electricity generation results in rapid oscillations of electricity injections into the grid, intermittent RES are putting stress on the infrastructure designed to guarantee stable flows of electricity. The Plan Wathelet also includes incentives for investments in new gas-fired capacity of 800 MW and incentives to increase demand side management (DSM) efforts equal to 400 MW by 2015 (on top of the existing DSM potential of 331 MW). The speed-up of interconnection plans with the U.K. and the Netherlands is also part of the Plan Wathelet There is a chance that one of the more recently built reactors lifetime will be extended in case of security of supply concerns. Furthermore, since March 2014 the two old nuclear reactors that were restarted in June 2013 have been closed again because of security concerns. 5

15 In our future scenarios - see below - we take the phase out as in the Plan Wathelet as our starting point. We also assume that new policy incentives will facilitate investments in gas plants (along with support for biomass and other renewables). Additionally, demand side management and interconnection will be mentioned in our analysis. The main focus will be on how security of supply can be achieved and on the main trade offs in terms of the overall costs for society. 2.3 The Central West European electricity market Belgium is a small country in the middle of Western Europe. By consequence, it is strongly influenced by the energy policies of its neighbouring countries. The interconnection capacity of Belgium is 3,5 GW, which is about 20% of peak demand. In other words, importing electricity can be a solution for a short term shortage. Unfortunately, France, Germany and Belgium have a limited production capacity to cope with very cold winters. The Winter Outlook report by ENTSO-E (2012) assigned code red to Belgium and code orange to France and Germany (Figure 6). Figure 6: Winter Outlook 2012 (ENTSO-E, 2012) Here we use the 2012 report (and not the 2013 outlook) because the situation in the Belgian electricity market in that winter was similar to the situation we will face in 2015, after the phase out of the oldest reactors. In 2012 there was a temporary shutdown of 2 nuclear reactors, resulting in a drop of 2 GW in 6

16 Reserve Margin (%) Johan Albrecht, Ruben Laleman the firm capacity. Also, the 2013/2014 winter was very warm and is not representative for a normal winter in Europe. The code red in the ENTSO-E report indicates that the peak demand in a strong winter would be higher than the total available capacity and total import capacity combined. In other words, a shortage in capacity can occur (resulting in blackouts). In Germany and France as well, some import would be needed under severe winter conditions, but the import capacity should be sufficient (in contrast to the Belgian case). The graph also shows that between Belgium and France, in case of a long and cold winter, there will no net-import or export, since they are both in the danger zone. France would import mostly from Spain (Red arrow). Belgium would mainly import electricity from the Netherlands (orange arrow from the Netherland to Belgium). A recent report from the CREG (CREG, 2013) presents historical data to support this, with strong import from the Netherlands in the winters of , and Import and export patterns between France and Belgium are much less seasonal. For example, there was a net-export situation from July 2009 until March In the winter of , there was net-export to France in October, November and January. The reserve margin for the CWE-region as a whole is about 20% (Figure 7). Only the Netherlands and Luxemburg have a RM above 25 percent. Despite the low reserve margin in Belgium, average electricity prices remain very low in Belgium because of interconnection with countries with relative overcapacity, an increasing influx of low-marginal cost renewable electricity and a sluggish demand. Too low prices do not trigger new investments and hence lead to even lower reserve margins in the next years... Figure 7: Reserve margins in CWE 4 (based on data from (ENTSO-E, 2012) (European Commission, 2013) (DENA, 2010)) 80% 70% 60% 50% 40% 30% 20% 10% 0% Rm + import Reserve margin BE DE FR LU NL CWE The reserve margin only takes into account the reliable or firm capacity. Total capacity did increase in recent years, but this was mainly due to increases in (non-reliable) wind and PV capacity. Also, cheap coal fired power plants in Germany are now determining market prices, resulting in CWE-prices of about 4 On the peak demand in Germany, very diverging data can be found; appendix III in the Commission report mentions a peak load of 92 GW, while DENA mentions data in the range of MW. A graph by IHS indicates a reserve margin of about 20% for Germany (IHS, 2013). We have opted for a peak demand of 83 MW in this study, since this seems to fit most of the literature estimates. 7

17 AverageWholesale Electricity Prices ( /MWh) Johan Albrecht, Ruben Laleman /MWh (Figure 8). A recent report from DG-Energy confirms this and refers to prices of /MWh for the second quarter of 2013 (European Commission, 2013). Figure 8: Average Wholesale prices in CWE (historic data based on (European Commission, 2004) (European Commission, 2006) (European Commission, 2008) (European Commission, 2012))

18 Reserve Margin (%) Johan Albrecht, Ruben Laleman 3 Future electricity supply scenarios for Belgium 3.1 Introduction Until 2020, some evolutions are rather predictable because they are driven by legislation. Firstly, there is the phase out of the two oldest nuclear reactors in Belgium, and the extension of Tihange 1. Secondly, we can expect a general increase in renewable electricity production in order to reach Belgium s national target of 20% renewable electricity production by 2020 (13% renewable energy in the whole economy). Based on these evolutions, we briefly evaluate a No Policy scenario as a benchmark to illustrate what would happen without incentives for additional investments in firm capacity. Then we evaluate several security of supply scenarios, to see which types of incentives are needed to guarantee a 5% reserve margin at all times in Belgium. Finally we evaluate the impact of the security of supply scenarios on the risk of oversupply and compare various renewable policy scenarios that can - or cannot - reduce the potential oversupply problem. For the No Policy scenario we combine the phase out plan with information about the decision to shut down or mothball existing capacity. By comparing generation capacity to the expected evolution of peak demand - we assume an increase by 0.5% per year (see 3.2.1) - the overall risk of shortages in Belgium is assessed. It is important to stress that we consider Belgium as an island. This simplifies the analysis and is considered to be the safest scenario. We also estimate the effect of 10% capacity credit or guaranteed contribution of wind energy to the reserve margin (Figure 9). The potential for PV to contribute to the reserve margin is close to zero because Belgium is a country with a peak demand during winter evenings. However, in Southern countries or regions (e.g. Spain, Italy) the contribution of solar technologies in response to peak demand can be very significant (EPIA, 2012). Figure 9: Reserve margin in the No Policy scenario 20% 10% 0% -10% -20% -30% -40% -50% -60% -70% -80% RM RM 10% CF wind Figure 9 clearly shows that the phase out of Doel 1 and 2, combined with the closing or mothballing of fossil assets (coal, gas) will result in a negative reserve margin from 2015 onwards. The situation will be 9

19 dramatic in with extremely low reserve margins in our No Policy scenario (without new investments). The above results clearly show that initiatives to increase the reserve margin are urgently needed. Therefore we will suggest security of supply alternatives in the following sections. Based on our findings from the No Policy scenario, we assess different policy options such as new incentives for flexible capacity, renewable generation and increasing the potential for demand side management (DSM). The different scenarios and assumptions are listed in Table 2. We aim to estimate overall costs of various combinations of these options, as well as the impact on the stability of the electricity system (by quantifying the size and frequency of surpluses). The details of the policy options will be discussed below. Table 2: Overview of policy options Shortage issues Surplus issues DSM potential Renewables incentives No DSM increase Contract For Difference (CFD) system 2100 MW additional DSM by 2030 CFD system + Market Participation (MP) Incentives for Flexible generation Support for new capacity only Support for New and Old Thermal capacity (OT) 3.2 Assumptions for all scenarios Evolution of peak demand First of all, we need an assessment of the evolution of peak demand. Our assessment is based on peak load data from Elia (see Figure 4) and on the FOD Economy Report on the supply of electricity in Belgium in (FOD Economy, 2012). In our study we assume a 0,5% increase of annual peak demand (starting at MW in 2013). This results in a peak demand of MW in However, demand side management could reduce this need (see Figure 16). The costs calculated in this report should be interpreted with these assumptions in mind. If we can decrease peak demand - by reducing energy consumption in winter or by the widespread use of DSM - costs in any given scenario will be lower. However, if peak demand increases, the opposite will be the case and overall costs will be higher Evolution of the carbon price The price of carbon in the ETS is currently at a very low level. This level certainly does not represent the real environmental cost of a ton of GHG-emissions. We therefore assume that in the future the price of carbon will increase: from about 8/ton CO 2-eq in 2014 to 40/ton CO 2-eq in Discount rate For the calculations of the levelized cost of electricity, we use a discount rate of 8%. Lower discount rates will result in lower costs. Higher discount rates will result in higher costs, especially for assets with high upfront investment costs (wind, PV, hydro and nuclear). 10

20 3.2.4 Cost calculations All costs are calculated in real terms and do not include inflation. We only consider cost changes resulting from our scenarios and estimates. The costs of the current system costs of using all assets, existing subsidies (e.g. for renewable generation),...- are not included in the subsidy costs of our scenarios. Negative cost changes can in principle occur, indicating that implementing a given policy will reduce the current system costs (e.g. by lowering the subsidy cost of renewable generation technologies). For example, in some scenarios the biomass assets are incentivized to run in a more flexible way. This will reduce the costs of support for biomass plants compared to the support scheme already in place. At the moment there is no real incentive for the flexible use of biomass capacity Technology assumptions For all evaluated technologies the assumptions on investment costs, load factor, learning rate and other parameters such as feedstock and operational and maintenance (O&M) costs can be found in Table 3. These assumptions are the baseline assumptions for Carbon costs will increase with time (see section 3.2.2) and investment costs will decease according to learning rates. Feedstock (coal, gas, uranium, biomass pellets) and O&M costs are assumed to be constant in the period concerned. This was done to improve transparency of the results. Predicting future feedstock costs is in any case highly speculative and very difficult. Table 3: Technology Properties (sources: (OECD - NEA, 2013) (EPIA, 2012) (Department of Energy and Climate Change, 2012) (IEA, 2010) (IEA, 2010) (Laleman, Balduccio, & Albrecht, 2012) Assumptions Life LF LR Invest. O&M Feedstock Carbon LCOE 2014 estimates time Cost Cost Cost Cost 2014 (years) (%) (%) ( /kw) ( /MWh) ( /MWh) ( /MWh) ( /MWh) PV Onshore Wind Offshore Wind Hydro BM (Large Scale) Coal CCGT Nuclear CHP OCGT The load factor is a special parameter as it depends on both the nature of the technology and also on market circumstances (prices, reserve margins in the CWE-region, weather patterns, import and export capacities, policy changes,... ). Especially the effect of policy changes on the load factor of various technologies will be discussed in more detail throughout this study. 11

21 Average wholesale electricity prices ( /MWh) Johan Albrecht, Ruben Laleman Endogenous price model In order to estimate the subsidy costs for RES with a support scheme responsive to market dynamics we assume a Contract-For-Difference approach (see 3.4.2). As this scheme responds to market prices, an estimate of the future wholesale electricity price is needed to assess the difference between the market price and the targeted support. It is obviously very difficult to extrapolate the price of electricity in the next 15 years. Our estimates are based on a combination of recent data by the EU commission (European Commission, 2013) and a report from OECD-NEA on the German electricity system with high shares of RES (OECD - NEA, 2013). The report by NEA/OECD finds an endogenous relationship between the share of RES and the wholesale market prices. Based on this relationship and recent wholesale prices, we linked wholesale prices to the share of renewables (RES). Figure 10 illustrates this relationship. The prices mentioned here are yearly averages. This is very important since there is a relatively strong variation in the day to day and hour to hour prices on wholesale electricity markets, depending on the availability of PV or wind assets and imports from neighbouring countries. This is important since CCGT s will only produce at market prices above marginal generation costs, i.e. when prices are around 50/MWh or higher. Figure 10: Relationship between average wholesale electricity prices and share of Renewables (blue data: (OECD - NEA, 2013)) Germany (OECD-NEA) Belgium (Our estimates) 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% Share RES Distribution and transmission cost estimates The cost of electricity supply is not limited to the sum of the costs of the generation assets. Costs in transmission and distribution should also be taken into account. The costs of connecting an offshore wind park to the grid are significantly higher than the costs of adding a CCGT plant. The OECD-NEA study of 2013 provides a relationship between the increase in RES and the increase in overall system costs. These findings are applied to our study to estimate total system costs in all scenarios. Figure 11 shows the relationship between the share of renewables and the additional distribution and transmission costs. It indicates that the share of transmission and distribution in the overall system costs increases to 14% if renewables increase to a share of 50% in overall electricity supply. 12

22 Installed Capacity of RES (MW) Share of Transmission and Dsitribution in System Cost (%) Johan Albrecht, Ruben Laleman Figure 11: RES share and transmission and distribution cost increase (based on data from (OECD - NEA, 2013)) 15% 10% 5% 0% 0% 10% 20% 30% 40% 50% 60% RES share 3.3 Security of supply scenarios In the secure supply scenarios we add a boundary condition that the electricity system should have a reserve margin (RM) above 5% at all times. This is a relatively strong assumption, given the strong decline in capacity in The 5% reserve margin can only be attained by adding new capacity. We focus on the costs of the incentives needed to motivate investors to install new capacity. To ensure a 5% RM at all times, our model autonomously triggers new investments once the RM is below 5% in a given year. We assume that every capacity shortage is filled with investments in CCGT, OCGT and biomass plants. We elaborate in section on the share of each technology in the additional supply to secure the 5% RM. In alternative scenarios, we evaluate the impact of an increase in DSM potential as well as a combination of incentives for new gas plants and support policies for old gas plants. Finally, a policy that combines all three options is discussed Business as usual (BAU) and high renewable scenario (High RES) For all policy options considered, we distinguish two possible evolutions of renewable capacity, namely a high renewable future and a business as usual renewable future. Both scenarios differ in the installed capacities of CCGT s, biomass, wind and PV that will be added to the system in the coming years. The evolution of biomass capacity on the one hand and PV and wind on the other hand is shown in Figure 12. Figure 12: Evolution of renewables installed capacity (biomass, wind, PV) in BAU and high Res scenarios High Res PV + Wind BAU RES PV + Wind High Res Biomass BAU RES Biomass 13

23 The difference between both renewable capacity scenarios is determined by policy choices regarding RES. Biomass is a special kind of RES, since it is not weather dependant. Additional biomass capacity contributes to a higher reserve margin. As a consequence, biomass can replace a share of the disappearing nuclear and fossil assets. This is why a jump in biomass capacity occurs in , since in that period, MW of nuclear capacity is planned to be phased out. The added capacity of biomass plants is higher in the High Res case and a lower in the BAU RES case. For offshore wind, we assumed in both renewable scenarios that all the approved Belgian offshore parks will be operational by 2023 resulting in a final offshore capacity of MW. PV and onshore wind are assumed to have a decreasing growth rate in the BAU RES scenario but a linear growth in the High Res scenario. For hydro we assume that no further growth is possible in Belgium, hence the installed capacity remains stable between today and The assumptions for the BAU and high renewables scenarios are summarized in Table 4. Table 4: Summary of assumptions for the BAU and High renewables scenarios In case of a shortage (RM < 5%), missing capacity is replaced with BAU RES High Res CCGT 60% 50% Biomass 20% 30% OCGT 20% 20% Final Capacity installed in 2030 (MW) BAU RES High Res Wind Onshore Wind Offshore PV Hydro Biomass TOTAL RES capacity Total installed capacity With respect to biomass, it is important to stress that additional capacity does not automatically result in more biomass electricity production. As will be explained later, biomass has the important advantage of being a controllable or firm generation technology that can be used in a flexible way in response to market signals. Depending on market circumstances, biomass can be used as a baseload plant with high load factors or as a flexible medium-load plant with much lower load factors. In short, the total amount of renewable electricity actually produced will strongly depend on the policies regarding biomass electricity production. 3.4 Incentive scenarios to guarantee security of supply Introduction Table 5 provides an overview of the possible policy scenarios. We combined two types of Renewables Supply policies (see next section) with three types of Capacity Need policies that can be combined 14

24 (the last Capacity Need scenario is a combination of the three previous ones) to obtain eight different scenarios. The details will be explained below. Table 5: Overview of Scenarios RES Supply Capacity Need NEW DSM Old Thermal DSM + OT CFD contract for difference CFD-MP CFD-market participation In all scenarios capacity payments are introduced to incentivize the build-up of controllable assets, i.e. we assume capacity payments for biomass plants, OCGTs and CCGTs. The capacity payments are part to the subsidy costs we will discuss later. Table 6 shows our assumptions on capacity payments. For biomass we opted for subsidizing only 50% of the upfront investment costs. The investment costs of biomass plants far exceed the investment costs of gas-powered plants but biomass plants also receive financial support per MWh of renewable electricity produced. We thus provide a hybrid incentive system for biomass; a capacity payment equal to 50% of the investment cost since biomass investments add to the reliable capacity, and a CFD-contract as a renewable support subsidy (see 3.4.2) covering all other costs (the remaining 50% of investment and all the operational costs). The support per MWh under this CFD-contract depends on the load factor (LF) of the biomass plant; the lower the LF, the higher the CFDsupport per MWh has to be to compensate higher capital costs per MWh produced. Gas fired plants only receive a capacity payment, and no separate incentives for production. As marginal costs for CCGT s are assumed be sufficiently low to trigger start up of power plants in case of rising demand. OCGT s have higher marginal costs but are assumed to only to produce electricity in times of scarcity. The latter situation is likely to occur more frequently in the coming decades as more intermittent renewables will be on line and more firm capacity is phased out. Table 6: Assumptions on Capacity Payments Technology Share of upfront investment covered Cap Payment ( / kw) Biomass 50% 1050 CCGT 100% 900 OCGT 100% 700 In our simplified model, we neglect the impact of new capacity additions on the profitability of existing capacity. One new CCGT implies less running hours for existing CCGT. This type of compression effect is not included in our calculations. We can however assume that the relevance of this compression effect will disappear once the nuclear phase out is completed. As a consequence, even a capacity remuneration equal to the upfront investment cost might not be attractive enough to trigger investments. Especially in a market outlook with high share of RES and hence low wholesale prices, potential investors in CCGT, OCGT and biomass capacity might ask a risk premium in addition to the upfront investment cost. This type of risk premium is not yet included in our model. Both factors suggest that our calculations of subsidy costs significantly underestimate the real subsidy cost to trigger new capacity. 15

25 3.4.2 Renewable supply measures Contract for Difference scenario (CFD) A CFD guarantees a minimal revenue stream for the power plant operator. In principle, the gap between the market price and the average electricity production costs is covered by the CFD-scheme. This system is currently suggested by the U.K. government to incentivize new nuclear power plants. Figure 13 provides a visualization of this concept. This system can also be applied to renewables. For wind and PV, the levelized cost of electricity is compared to the estimated average electricity price on the market, and thus we can estimate the cost of the CFD-policy. For biomass, the LCOE is calculated based on an investment cost of 1050/kW (50% of investment, since it also receives a capacity payment, see previous paragraph). Figure 13: Visualization of CFD concept (own illustration) Market Participation (CFD-MP) In the CFD scenario, renewables are not incentivized to reduce output in times of low demand with a potential risk for overproduction (e.g. on a sunny summer weekend with strong wind). In the CFD-MP scenario we still assume a CFD-mechanism, but add some constraints on the feed-in of variable renewable electricity production (wind and PV). Also, we assume that biomass will be operated in a flexible way to respond to market needs. As a consequence, biomass plants run with a load factor of only 35% (about half the current load factor) while PV and wind can be curtailed under specific circumstances. We assume that PV and wind are only curtailed at times when they produce above 50% of their theoretical maximal output - from a country wide perspective. Beware, this does not entail that wind and PV will always be curtailed at times when their load factors exceed 50%. This assumption simply implies that wind and PV will never be curtailed when they produce less than 50% of their 16

26 theoretical maximal output 5. Based on real production data from , we concluded that curtailments will be rather rare. Figure 14 illustrates that the total output of all PV panels in Belgium rarely exceeds 50% of theoretical maximal capacity of all PV-systems installed in Belgium. This is actually not that surprising in a country with modest solar conditions. If we apply the market participation policy, this would result in a rather modest decline of the load factor of PV from 10,8% to 10,2%. From the graph we can see that only 5% of the time it would be necessary to actually top-off some PV-systems. In reality it might be that these peaks of over 50% occur at times with little wind or high electricity demand so that in practice curtailment is not always required in these cases. Curtailment of residential PV installations requires the roll-out of a smart grid with smart meters. We realize that there will not be such a comprehensive smart grid in Belgium in the next years. Nevertheless, a further increase of residential PV capacity without the technical ability to control and regulate PV output is not a consistent policy scenario. Up to 2030 and beyond, a smart grid is needed to facilitate the further expansion of residential PV capacity. Figure 14: Probability curves for the load factor of PV-production in Belgium (Elia, 2013) The same idea is applied to wind energy. However, the probability of the total wind output in Belgium going above 50% of nominal output is slightly higher (Figure 15). This is not really an issue since it would probably be easier to curtail wind production compared to PV production. 5 Let s clarify this curtailment option with a numerical example. If, in a given year, the total capacity of PV-systems in Belgium is equal to 6000 MW, curtailment will be possible once the production of all the PV-systems combined exceed 3000 MW. The same goes for wind capacity. 17

27 Figure 15: Probability curves for the load factor of onshore wind production in Belgium (Elia, 2013) Capping the total wind production in Belgium at 50% of theoretical maximum output would result in a decrease of total production of 14%. However, when wind would be above this threshold at times of high demand, it will not be curtailed. It is only considered to be possible to do so, if needed. Actually, in practice there is already some load shedding of wind occurring in Belgium at times of low demand and high wind speeds. For offshore wind insufficient data is currently available. Therefore we assume a similar (14%) decrease in output when capping at 50% of nominal capacity. Summary of Renewable supply policy options In all scenarios, the CFD incentive scheme is applied to wind (onshore and offshore), PV and biomass. It is designed in such a way that whatever the design of the system -with or without market participationthe total returns for the investors will be equal to the LCOE of the technology. Because of this, the LCOE for the technologies will be slightly larger in the CFD-Market Participation scenario, since the lower load factors need to be compensated by increasing the CFD-payments. Biomass is a special case; the CFD cost is based on only a part of the overall costs, since biomass investments also receive a capacity payment (see Table 6). Also, the load factor for biomass assets decreases the most (from 60% to 35%) because biomass plants have the ability to follow the load and can be used in a flexible way Incentives for firm capacity The Plan Wathelet specifically mentions some incentives to stimulate new capacity additions on top of the increase of demand side management and an extension of the strategic reserve as options for dealing with the supply concerns. Given the nuclear phase out and the current lack of market incentives for new capacity, this is a very legitimate policy concern. We will evaluate these options in this study. 18